Journal of Proteomics - University of Liverpool · Proteomics encompasses the comprehensive...

Contents lists available at ScienceDirect

Journal of Proteomics

journal homepage: www.elsevier.com/locate/jprot

The synovial fluid proteome differentiates between septic and nonsepticarticular pathologies

James R. Andersona, Aibek Smagula, Deborah Simpsonb, Peter D. Clegga,Luis M. Rubio-Martinezc,1, Mandy J. Peffersa,⁎,1

a Institute of Ageing and Chronic Disease, William Henry Duncan Building, 6 West Derby Street, Liverpool L7 8TX, UKb Centre for Proteome Research, Institute of Integrative Biology, Biosciences Building, Crown Street, University of Liverpool, Liverpool L69 7ZB, UKc Department of Equine Clinical Studies, Institute of Veterinary Science, Chester High Road, Neston CH64 7TE, UK

A R T I C L E I N F O

Keywords:Synovial fluidArthropathyOsteochondrosisOsteoarthritisSynovial sepsis,110 equine

A B S T R A C T

Articular conditions are common in horses and can result in loss of function, chronic pain and/or inability towork. Common conditions include osteoarthritis, osteochondrosis and synovial sepsis, which can be life-threa-tening, but despite the high clinical prevalence of these conditions, rapid and specific diagnosis, monitoring andprognostication remains a challenge for practicing veterinarians. Synovial fluid from a range of arthropathieswas enriched for low abundance proteins using combinatorial peptide ligand ProteoMiner™ beads and analysedvia liquid chromatography-tandem mass spectrometry. Changes in protein abundances were analysed usinglabel-free quantification. Principle component analysis of differentially expressed proteins identified groupingsassociated with joint pathology. Findings were validated using ELISA. Lactotransferrin (LTF) abundance wasincreased in sepsis compared to all other groups and insulin-like growth factor-binding protein 6 (IGFBP6)abundance decreased in sepsis compared to other disease groups. Pathway analysis identified upregulation of thecomplement system in synovial joint sepsis and the downregulation of eukaryotic translation initiation factorsand mTOR signalling pathways in both OA and OC compared to the healthy group. Overall, we have identified acatalogue of proteins which we propose to be involved in osteoarthritis, osteochondrosis and synovial sepsispathogenesis.Significance: Osteoarthritis, osteochondrosis and synovial sepsis, which can be life-threatening, are commonarticular conditions in which rapid and specific diagnosis, monitoring and prognostication remains a challengefor practicing veterinarians. This study has identified that the equine synovial fluid proteome exhibits distinctiveprofile changes between osteoarthritis, osteochondrosis, synovial sepsis and healthy joints. Elevated synovialabundance of lactotransferrin and decreased insulin-like growth factor-binding protein 6 were both found todistinguish synovial sepsis from all other study groups. Thus, these protein markers may have a future role inclinical practice to enable an earlier and reliable diagnosis of synovial sepsis.

1. Introduction

Articular conditions are common in horses and can result in loss offunction, chronic pain and/or inability to work, which are importantwelfare concerns. The most common conditions include osteoarthritis(OA), osteochondrosis (OC) and synovial sepsis, which can be life-threatening [1]. Despite the high clinical prevalence of these condi-tions, rapid and specific diagnosis, monitoring and prognostication re-mains a challenge for practicing veterinarians, thus, identification ofreliable biomarkers of disease is required. Synovial fluid (SF) is in direct

contact with articular structures and represents an important source ofbiomarker discovery. SF is located within the articular joint cavity,providing a pool of nutrients for surrounding tissues but primarilyserving as a biological lubricant, containing molecules including hya-luronan with low-friction and low-wear properties to articular surfaces[2]. As SF is in close proximity to articular tissues primarily alteredduring joint pathology, this biofluid is an important first approachsource for biomarker discovery [3,4].

Total protein content in synovial fluid is associated with presence ofarticular pathology and is one of the parameters most commonly used

https://doi.org/10.1016/j.jprot.2019.04.020Received 7 January 2019; Received in revised form 28 March 2019; Accepted 22 April 2019

⁎ Corresponding author.E-mail addresses: [email protected] (J.R. Anderson), [email protected] (A. Smagul), [email protected] (D. Simpson),

[email protected] (P.D. Clegg), [email protected] (L.M. Rubio-Martinez), [email protected], [email protected] (M.J. Peffers).1 Joint last author.

Journal of Proteomics 202 (2019) 103370

Available online 25 April 20191874-3919/ © 2019 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY license (http://creativecommons.org/licenses/BY/4.0/).

T

http://www.sciencedirect.com/science/journal/18743919https://www.elsevier.com/locate/jprothttps://doi.org/10.1016/j.jprot.2019.04.020https://doi.org/10.1016/j.jprot.2019.04.020mailto:[email protected]:[email protected]:[email protected]:[email protected]:[email protected]:[email protected]:[email protected]://doi.org/10.1016/j.jprot.2019.04.020http://crossmark.crossref.org/dialog/?doi=10.1016/j.jprot.2019.04.020&domain=pdf

in equine clinical practice to enable diagnosis and to monitor horseswith synovial sepsis [5]. Problematically there are large overlaps in theranges of SF total protein content among arthropathies, and total pro-tein content is also affected by articular diagnostic procedures andtreatments [6]. This situation can lead to erroneous interpretations andclinical decisions, representing an important welfare risk to the horse.More specific proteins such as the acute phase inflammatory proteinserum amyloid A has been investigated as a potential biomarker ofarticular conditions in horses [7]; however, it is easily affected by thesystemic condition of the patient, and therefore, can be unreliable forspecific clinical diagnosis [8].

Proteomics encompasses the comprehensive profiling of proteincontents. The proteomic profile of the synovial fluid is dependent uponarticular disease and its characterization has allowed the differentiationof OA and rheumatoid arthritis (RA) in man [4]. A recent study hasidentified a subset of proteins differentially expressed in OA versusnormal equine joints [9]. Although an increase in total protein is a well-known feature of synovial sepsis, the protein profile in synovial sepsishas not been investigated. Identification of a characteristic proteomicprofile in synovial sepsis would increase the understanding of the mo-lecular basis of the condition and also enable the identification of po-tential biomarkers for early diagnosis and treatment of this disease.

A previous study has investigated the equine SF profile in OA andOC compared to healthy donors using 2D-gel electrophoresis withsubsequent mass spectrometry analysis of in-gel tryptic digests [46].However, few studies have investigated the proteomic profile of equineSF using high mass resolution mass spectrometer of in-solution trypticdigestions [9,10] and, to our knowledge, none have investigated pro-tein profile changes in septic SF in any species. In this study we haveused mass spectrometry and label-free quantification to identify po-tential protein biomarkers that allow differentiation between equinearticular joint pathologies, potentially aiding accurate diagnosis andenabling increased understanding of the effect of septic synovitis on thejoint environment. Identification of characteristic proteomic profiles insynovial sepsis would increase the understanding of the molecular basisof the condition and also allow the determination of potential bio-markers for early diagnosis and treatment of the disease. We hy-pothesised that proteomic profile characteristics will distinguish septicfrom non-septic equine arthropathies and between the non-septiccommon arthropathies OC and OA. The objective of the study was toidentify possible synovial biomarkers in equine arthropathies.

2. Methods

All chemicals were supplied by Sigma-Aldrich (Gillingham, UK)unless otherwise stated.

2.1. Sample collection

Following ethical approval and owner consent, excess aspirated SF(collected during clinical diagnostic investigations) was analysed fromjoints of horses presenting to The Philip Leverhulme Equine Hospital,University of Liverpool between 2014 and 2016, diagnosed with clin-ical OA, OC and synovial sepsis. Clinical diagnoses were determined viaa combination of radiography, ultrasonography, arthroscopy and SFanalysis and/or bacterial culture as previously described [11]. SF wasaspirated for diagnostic purposes from the affected joints during patientexaminations or at the start of surgical arthroscopy under general an-aesthesia. SF was immediately placed into EDTA-containing tubes (BDVacutainer, Belliver Industrial Estate, Plymouth, UK) and particulateand cells were removed from the SF by centrifugation (4 °C, 2540 g for4min). The cell-free supernatant was transferred to a clean uncoated1.5 ml collection tube, snap-frozen using liquid nitrogen and stored at-80 °C. SF was collected from the distal interphalangeal, femorotibial,glenohumeral, intercarpal, metacarpophalangeal, metatarsophalangeal,patellofemoral and tarsocrural joints.

Eight healthy synovial fluid (SF) samples were obtained from themetacarpophalangeal joint of mixed breeds of horses from an abattoirwith no gross joint changes, with a total score of 0 according to theMcilwraith et al. gross scoring system, scored by two independent as-sessors [12]. Samples were collected as a by-product of the agriculturalindustry and processed within 12 h of euthanasia. The Animals (Sci-entific Procedures) Act 1986, Schedule 2, does not define collectionfrom these sources as scientific procedures and ethical approval wastherefore not required. Post-mortem SF samples were processed usingidentical protocols as for clinically collected hospital samples. Finalgroups consisted of OA (n=4), OC (n=8), synovial sepsis (n=7) andhealthy (n= 8) (full details in Supplementary File 1).

2.2. Synovial fluid preparation

SF was thawed and treated with 1 μg/ml hyaluronidase as pre-viously described [9] and centrifuged at 10000 g for 10min to removeany particulates. Protein concentrations of SF were determined byBradford assay (Thermo Scientific, USA). A volume of sample equiva-lent to 5mg of protein was added to 10 μl of beads [9]. SF samples wereenriched for low abundance proteins using ProteoMiner™ beads(BioRad, UK) according to the manufacturer's instructions. Beads werere-suspended in 80 μl of 25mM ammonium bicarbonate and 5 μl of 1%(w/v) Rapigest SF (Waters, UK) added and the sample heated at 80 °Cfor 10min. On bead trypsin digestion was undertaken [9].

2.3. Liquid chromatography tandem mass spectrometry (LC-MS/MS) andlabel-free quantification

Tryptic digests (samples were diluted 5-fold in 0.1% (v/v) TFA/3%(v/v) acetonitrile and 2 μl loaded) were subjected to LC-MS/MS, using a2 h gradient. Data-dependent LC-MS/MS analyses were conducted on aQExactive HF quadrupole-Orbitrap mass spectrometer (ThermoScientific, Hemel Hempstead, UK) coupled to a Dionex Ultimate 3000RSLC nano-liquid chromatograph (Thermo Scientific). Sample digestswere loaded onto a trapping column (Acclaim PepMap 100 C18,75 μm×2 cm, 3 μm packing material, 100 Å) using a loading buffer of0.1% (v/v) trifluoroacetic acid, 2% (v/v) acetonitrile in water for 7minat a flow rate of 12 μl min−1. The trapping column was then set in-linewith an analytical column (EASY-Spray PepMap RSLC C18,

Fig. 1. Synovial fluid protein concentrations in equine arthropathies. Proteinconcentrations of SF were determined by Bradford assay. Healthy (n=8), OA(n=4), OC (n=8), sepsis (n=7). Significant changes are demonstrated fol-lowing analysis using ANOVA with Tukey's multiple comparison; ** p < .01,**** p < .0001.

J.R. Anderson, et al. Journal of Proteomics 202 (2019) 103370

2

75 μm×50 cm, 2 μm packing material, 100 Å) and the peptides elutedusing a linear gradient of 96.2% A (0.1% [v/v] formic acid):3.8% B(0.1% [v/v[formic acid in water:acetonitrile [80,20] [v/v]) to 50%A:50% B over 90min at a flow rate of 300 nl min−1, followed bywashing at 1% A:99% B for 5min and re-equilibration of the column tostarting conditions. The column was maintained at 40 °C, and the ef-fluent introduced directly into the integrated nano-electrospray ioni-sation source operating in positive ion mode. The mass spectrometerwas operated in DDA mode with survey scans between m/z 350–2000acquired at a mass resolution of 60,000 (FWHM) at m/z 200. Themaximum injection time was 100ms, and the automatic gain controlwas set to 3e6. The 12 most intense precursor ions with charges statesof between 2+ and 5+ were selected for MS/MS with an isolationwindow of 2m/z units. The maximum injection time was 100ms, andthe automatic gain control was set to 1e5. Fragmentation of the peptideswas by higher-energy collisional dissociation using normalised collisionenergy of 30%. Dynamic exclusion of m/z values to prevent repeatedfragmentation of the same peptide was used with an exclusion time of

20 s.For label-free quantification, the raw files of the acquired spectra

were analysed by the ProgenesisQI™ software (Waters, Manchester, UK)[9] which aligns the files and then peak picks for quantification bypeptide abundance. Briefly, the top five spectra for each feature wereexported from ProgenesisQI™ and utilised for peptide identificationwith our local Mascot server (Version 2.6.2), searching against theUnihorse database with carbamidomethyl cysteine as a fixed mod-ification and methionine oxidation as a variable mod, peptide masstolerance of 10 ppm and MSMS tolerance of 0.01 Da. In this study wedefine differentially expressed (DE) proteins as those with a> 2 foldchange in expression, with a false discovery rate (FDR) adjusted p valueof< 0.05, and identified with at least two unique peptides.

Proteomic data has been deposited in the PRIDE ProteomeXchangeand can be accessed using the identifier PXD011276 [13].

Fig. 2. Analysis of synovial fluid proteomics. (A) Venn diagram of protein identifications for each analysed group. A minimum of 2 unique peptides were required forprotein identification, with a 1% FDR correction applied. (B) Principal component analysis plots of healthy versus OA, healthy versus OC and healthy versus sepsis(FDR < 0.05,> 2-fold change and identified with at least two unique peptides).


3

2.4. Pathway and network analysis of proteomic data in contrasts of eithersepsis, OA or OC versus healthy

Network analysis, canonical pathways and upregulated regulators ofproteomics data were analysed through the use of IPA (Qiagen, US).Proteins with an FDR adjusted p value of< 0.05 and with 2 or moreunique peptides were used for analysis, uncharacterised proteins werenot included to the analysis. Gene names from the lists generated by

ProgenesisQI™ were uploaded on IPA, unmapped genes were convertedmanually to human orthologues and as species mammals were checkedfor analysis.

2.5. Validation of mass spectrometry using ELISA

The differentially expressed proteins lactotransferrin (LTF) and in-sulin-like growth factor-binding protein 6 (IGFBP6) were selected tovalidate mass spectrometry findings as they could be measured usingavailable enzyme linked immunosorbent assays (ELISAs) that werecompatible with equine samples. Commercially available kits were usedfor both equine LTF (MBS902183, MyBioSource Inc., San Diego,California, USA) and equine IGFBP6 (abx574245, Abbexa Ltd.,Cambridge, UK), using sandwich and competitive inhibition ELISAmethodology respectively. For IGFBP6 analysis, HA treated/Costarprocessed native SF was diluted 1/10 whereas for LTF analysis SF wasundiluted. 5–6 dependent cohort SF samples were tested per group.100 μl aliquots of each sample were analysed in duplicate, with theabsorbance measured at 450 nm and protein concentrations calculated

Fig. 3. Principal component analysis and heat map analyses of healthy, OA, OC and sepsis SF samples. (A) Principal component analysis of proteins(FDR < 0.05,> 2-fold change and identified with at least two unique peptides) revealed the greatest variability was due to the type of arthropathy distinguished byshading (healthy; red, OA; green, OC; blue, sepsis; turquoise). (B) Heat maps of the top 100 differentially expressed proteins show distinguishable profiles based onarthropathy type. Each column indicates a different sample group. Each row indicates one protein. Red shading indicates increased expression and blue decreasedexpression. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Table 1Number of proteins DE in two-way comparisons.

Contrast Number of Proteins DE Increased Decreased

Healthy versus Sepsis 386 205 181Healthy versus OA 430 331 99Healthy versus OC 539 474 65OA versus Sepsis 53 22 31OC versus Sepsis 454 75 379OA versus OC 0 0 0


4

from standard curves.

2.6. Additional statistical analyses

Principal component analysis (PCA) plots and heat map analysiswere produced using MetaboAnalyst 4.0. The Venn diagram was pro-duced using the online tool Venny 2.1 and box plots constructed usingGraphPad Prism 8 and SPSS 24. Statistical analysis of SF protein con-centration, LTF and IGFBP6 abundances were performed in Minitabversion 17. These were conducted via ANOVA, using Tukey post hoctesting, with a p value of< 0.05 considered statistically significant,following correction for multiple testing.

3. Results

3.1. Protein concentration of SF different arthropathies

The SF protein assay demonstrated that concentration was de-pendant on the type of arthropathy (Fig. 1).

3.2. Differential abundance of proteins in healthy compared toosteoarthritis, osteochondrosis and sepsis SF

From the aggregate data file produced from all samples 1385 pro-tein families, at a 1% FDR, were identified (Supplementary File 2).ProgenesisQI™ identified 750 proteins DE when all groups were com-pared using a 4-way analysis (> 2-fold change, FDR < 0.05 andidentified with at least two unique peptides) (Supplementary File 3).The number of protein identifications in each analysed group are shownin Fig. 2A. The number of DE proteins in 2-way contrasts is shown inTable 1 (with complete protein lists shown in Supplementary File 4)and multivariate 2-way comparisons of protein profiles shown inFig. 2B. Unsupervised multivariate analysis (Principal ComponentAnalysis (PCA)) demonstrated clear variance between groups in the 4-way analysis (Fig. 3A).

3.3. DE proteins were filtered to include proteins identified with two or morepeptides, q < 0.05 and > 2 fold change in expression

Canonical pathways and network analysis of osteoarthritis, osteochon-drosis and sepsis SF.

Canonical pathway analysis was undertaken on differentiallyabundant proteins, contrasting either OA, OC or sepsis, versus healthy(Supplementary File 5). When sepsis samples were compared to healthyit was revealed that the predicted upregulation of coagulation system(p=7.94× 10−23), acute phase response signalling(p=3.98× 10−19), LXR/RXR activation (p=2.51×10−17), extrinsicand intrinsic prothrombin activation pathways (p= 2×10−16 andp=1×10−14), complement system (p=3.16×10−11), neuropro-tective role of THOP1 in Alzheimer's disease (p=1.55× 10−6) and OApathway (p=7.08× 10−6); and inhibition of matrix metalloproteaseswas downregulated (p= 2.51× 10–15) (Fig. 4A). Whereas pathwayanalysis of OA versus healthy demonstrated upregulation of the coa-gulation system (p=5.01×10−17), complement system(p=1.58× 10−16), extrinsic and intrinsic prothrombin activationpathways (p=5.01×10−12 and p= 2×10−11) (Fig. 4B). Compar-ison of differential abundant proteins in OC compared to healthysamples demonstrated an upregulation in complement system(p= 2×10−16) and RhoGDI signalling (p=5.62×10−9) (Fig. 4C).Downregulated canonical pathways of OA or OC versus healthy had asubstantial overlap including acute phase response (OAp=6.31× 10−15, OC p=7.94× 10−11), EIF2 signalling (OAp=7.94×10−15, OC p=1.26× 10−23), actin cytoskeleton signalling(OA p=7.94×10−13, OC p=1×10−12), regulation of eIF4 andp7056K signalling (OA p=7.94×10−13, OC p=3.98× 10−15),mTOR signalling (OA p=1.58× 10−11, OC p=3.16×10–13), leu-kocyte extravasation signalling (OA p=3.98× 10−11, OCp=3.98×10−11), signalling by Rho family GTPases (OAp=2.82× 10−10, OC p=1.82×10−9), p7056K signalling (OAp=8.91× 10−9, OC p=1.15×10−9) and protein kinase A signalling(OA p=1.02× 10−8, OC p=4.9×10−8). IGF-1 signalling(p= 3.16× 10−11) and ephrin B signalling (p=1.91×10−9) were

Fig. 4. Canonical pathways of differentially abundant proteins in equine synovial fluid of (A) sepsis, (B) OA and (C) OC compared to healthy samples. Bars representthe significance of the canonical pathway, calculated by a right-sided Fisher's exact test. The tallest bars represent the canonical pathways that are the least likely tohave been identified due to random chance. Upregulated canonical pathways are shown in orange and downregulated pathways in blue. (For interpretation of thereferences to colour in this figure legend, the reader is referred to the web version of this article.)


5

downregulated pathways in OA only; and Fcy receptor-moderatedphagocytosis in macrophages and monocytes (p=1.91×10−10),production of nitric oxide and reactive oxygen species in macrophages(p=2.88×10−8) present in OC only.

Compared to healthy SF, in sepsis there was an upregulation incellular movement, haematological system development, inflammatoryresponse, cell-to-cell signalling and immune cell trafficking and this wasin distinct contrast to these pathways in OA and OC in which they weredownregulated (Fig. 5A). Similarities between groupings of canonicalpathways in OA (Fig. 5B) and OC (Fig. 5C) were identified. The groupsof organismal injury and abnormalities, and cancer had the highestoverlap in healthy versus sepsis/OA/OC, however the activation ofmost pathways could not be predicted by IPA, with an overall trend tobe downregulated. Organismal injury and cancer had the most overlapin healthy versus OA or OC as well. These were predicted to be upre-gulated in OA compared to sepsis and even more upregulated in OC.(See Fig. 6.)

ELISA validation of lactotransferrin and Insulin-like growth factor-binding protein 6.

LTF ELISA assay results corroborated mass spectrometry findingswith a statistically elevated abundance present within sepsis SF samplescompared to all other groups (Fig. 7). IGFBP6 ELISA results produced a

similar trend to that identified via mass spectrometry; however theseresults were not of statistical significance.

4. Discussion

Articular pathologies have an important clinical relevance in horses;however, to date the veterinary practitioner can find it difficult toachieve a timely diagnosis and prognostication of these conditions inhorses affected by joint pathologies. This study has confirmed that SFprotein concentration is associated with the type of arthropathy and hasidentified pathways and possible markers that may enable further un-derstanding, together with rapid and accurate identification of theseconditions in horses and potentially other species. For the first time thestudy of the proteomic profile of SF has identified potential SF markersof synovial sepsis. The proteomic profile can discriminate betweenseptic and non-septic articular pathologies, and between non-septicconditions such as OC and OA. The implications highlighted by proteinchanges attributable to different pathways has increased our knowledgeon articular pathophysiology.

Within this study, elevated synovial levels of lactotransferrin (LTF)were identified within the sepsis group compared to the other pathol-ogies. LTF, also known as lactoferrin, is an 80 kDa multifunctional

Fig. 5. Heat map identifying canonical pathway groupings for molecular and cellular functions altered in equine synovial fluid of (A) sepsis, (B) OA and (C) OC incomparison to healthy samples. Squares are coloured according to their Z score, with orange upregulated in the disease state and blue downregulated, with the colourintensity indicating prediction strength. The Z score represents whether the up- or downregulation of the proteins within that function will lead to activation (positiveZ score) or inhibition (negative Z score) of the function. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version ofthis article.)


6

glycoprotein, abundant in various biofluids, primarily transporting ironions, but also exhibiting anti-microbial, anti-inflammatory, and im-mune modulatory effects [14–16] LTF provides anti-bacterial actionagainst various bacterial species, inhibiting the growth of Staphylo-coccus epidermidis, antibiotic-resistant Klebseilla and Staphylococcusaureus in vitro and reducing blood and hepatic infection severity fol-lowing Escherichia coli enteral infection in rats [17–20] LTF is presentwithin specific granules of neutrophils with elevations in LTF levelspotentially representing a marker of neutrophil granulocyte activation[21]. LTF levels within human SF have been found to be closely cor-related to the level of SF neutrophilia, indicating the degree of in-flammatory response [22]. Various studies on SF from RA patients haveidentified increased synovial levels of LTF [23–25]. LTF has been foundto activate expression of bone morphogenetic protein 7 (BMP7) withinporcine articular chondrocytes, through the mitogen-activated proteinkinase ERK pathway [26]. As BMP7 has an important role in main-taining homeostasis of articular cartilage, the authors proposed LTFexhibits a protective function following inflammation within the joint.During synovial sepsis, there is an upregulation of collagen catabolismdue to an increased cytokine concentration leading to the release ofmatrix metalloproteinases [27]. Thus, following synovial sepsis therelease of LTF from neutrophil granulocytes may have a multi-func-tional role, both providing a direct antibacterial function and enabling adegree of chondroprotection. Clinically, LTF diet supplementation ofneonates has demonstrated an ability to reduce the occurrence of late

onset sepsis [19,28–30]. However, elevations in synovial LTF levelshave not previously been identified as a marker of synovial sepsis.

In this study, reduced SF IGFBP6 was identified within the sepsisgroup compared to other joint pathologies. This protein is involved incartilage and bone homeostasis but is also expressed in the synovium,particularly in RA versus OA [31]. IGFBP6 is a 30 kDa protein which issecreted into extracellular regions where it can subsequently interactwith insulin growth factors (IGFs). IGFBP6, unlike other IGFBPs, bindsto IGF2 with higher affinity than IGF1, inhibiting IGF2 activity [32].IGFBP6 has previously been identified in human SF, with synovial le-vels decreased in RA compared to OA [31]. Postoperative sepsis sig-nificantly alters the growth hormone/IGF axis, reducing GH-dependentmolecule secretion (i.e. IGFBP3) and increasing GH-independent mo-lecule secretion (i.e. IGFBP6) [33]. However, decreased levels of sy-novial IGFBP6 have not previously been identified for synovial sepsis inhorses. IGFBPs are known regulators of insulin resistance, with in-creased serum IGFBP6 levels identified within type 1 diabetes, and thushigh systemic glucose [34]. Reduced synovial glucose concentrations isa known clinical parameter associated with synovial sepsis [35]. Thus,the localised reduced abundance of IGFBP6 may be reflective or thedecreased synovial glucose levels.

Protein differences were greatest between synovial sepsis and othergroups probably reflecting the more distant pathological pathwaysimplicated. Pathway analysis of DE proteins in sepsis compared tohealthy revealed upregulation of a number of canonical pathways

Fig. 6. Network involved in the immune response of cells mapped from the differential abundant proteins of septic equine synovial fluid. Red nodes represent greaterprotein abundance in sepsis; green nodes represent lower protein abundance in sepsis; and white nodes represent molecules mapped to this network and not presentin the dataset. Orange arrows represent predicted activation, blue arrows predicted inhibition and yellow arrows indicate results inconsistent with the state ofdownstream molecules. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)


7

including the coagulation system, acute phase response signalling andcomplement system. Systemic inflammation results in activation ofcoagulation, due to tissue factor-mediated thrombin generation,downregulation of physiological anticoagulant mechanisms, and in-hibition of fibrinolysis [36]. In healthy SF, coagulation proteins have alow abundance with the elevated SF levels identified within sepsis inour study probably due to thrombin activation and coagulation, alsoreported in inflamed human joints [37]. In an equine model of jointinflammation changes in fibrinogen and thrombin/antithrombin havealso been reported [38]. The upregulation in sepsis of the acute phaseresponse signalling pathway is supports findings of our group andothers that acute phase proteins such as serum amyloid A increase inequine joint sepsis [11,39].

The upregulation of the complement system in joint sepsis is notunexpected. The complement system is one of the key players in thedefence against infections [40]. Its activation during the innate immuneresponse leads to the generation of several proteins that contribute tothe lysis and opsonisation of microorganisms, regulate inflammatoryreactions and bridge innate immunity with the subsequent adaptiveimmune response. Depletion in the complement system in animalmodels increases the severity of septic arthritis [41].

Interestingly eukaryotic translation initiation factors (EIF2) andmTOR signalling were downregulated pathways in both OA and OCcompared to healthy SF. OC is a juvenile osteoarticular disorder af-fecting several mammalian species. In horses, it is a multifactorialdisease with focal disruption of endochondral ossification leading to thedevelopment of osteoarticular lesions [42]. Nevertheless, OC patho-physiology is poorly understood. Protein synthesis requires cooperationamong a large number of polypeptides including ribosomal proteins,modification enzymes and ribosome associated translation factors. Theinitiation phase of protein synthesis requires a set of EIFs. A reductionin these proteins in OA and OC is interesting as in both conditions thereis an imbalance between cartilage protein anabolism and catabolismleading to a net loss of cartilage. However there is an increase in mTORin end-stage OA and mTOR is important in maintaining cartilagehomeostasis [43]. Furthermore signalling through mTOR is an im-portant link between synovitis and structural damage in inflammatoryarthritis [44]. In OA and OC proteins in the SF in our study may bederived from cartilage, subchondral bone or synovium. Further studiesare required to determine the relevance of these changes at the SFproteome level to the pathogenesis of these diseases since mTOR in-hibition protects cartilage from experimental OA (42).

Fig. 7. ELISA validation of LC-MS/MS results for lactotransferrin (A and B) and insulin-like growth factor binding protein 6 (IGFBP6) (C and D). IGFBP6 normalisedto total protein. ELISA: n=5–6/group. *= p < .05, **= p < .01, ***= p < .001. OA=osteoarthritis, OC= osteochondrosis, n.s. = non-significant.


8

Using IPA we have identified a number of proteins requiring furtherinvestigation as potential biomarkers of disease and/or therapeutictargets (Supplementary File 5). For example, Nuclear factor (erythroid-derived 2) (NFE2L2) was identified as the most significantly inhibitedupstream regulator in OC. Interestingly although also identified as apotential upstream inhibitory regulator in OA the effect in OC wasmuch more significant. This master regulator of redox homeostasisregulates the expression of antioxidant proteins that protect againstoxidative damage triggered by injury and inflammation. It has a po-tentially protective role in joint inflammation and degeneration [45].We have therefore identified a catalogue of proteins for future studies.

Healthy SF samples were collected following euthanasia whilstsamples from the pathological groups were predominantly collectedfrom live horses during clinical examination. Thus, it cannot be ruledout that differentially abundant proteins identified within the healthygroup may be resultant of post mortem change. However, post mortemprocessing took place within 12 h of euthanasia and therefore the im-pact on the synovial proteome is likely to be minimal. Due to the lim-ited sample sizes included within this study, SF aspirated from variousjoints was included. However, despite the inclusion of analysis of SFfrom multiple sites, the clear separation identified in the proteomicprofiles of different joint diseases strengthens the conclusion that thedifferentially expressed proteins are resultant of the joint pathologyopposed to joint location. Additionally, horses with different patholo-gies investigated within this study may experience changes in man-agement, including diet and exercise, which may subsequently influ-ence synovial protein composition. However, these systemic effects arelikely to have a limited effect on the synovial proteome compared to themore significant impact of local joint pathology.

Within this study, ELISA validations of LTF and IGFBP6 synovialabundances were undertaken on the same samples as those used formass spectrometry analysis due to limited availability of additional SFsamples. Thus, to confirm the results of this study identifying these aspotential synovial sepsis markers, further validation needs to be un-dertaken using an independent cohort before these proteins are in-vestigated further.

5. Conclusion

This study has identified that the equine SF proteome exhibits dis-tinctive profile changes between OA, OC, synovial sepsis and healthyjoints. Elevated synovial abundance of LTF and decreased IGFBP6 wereboth found to distinguish synovial sepsis from all other study groups.Thus, these protein markers may have a future role in clinical practiceto enable an earlier and reliable diagnosis of synovial sepsis.

Acknowledgements

The authors would like to thank The Centre for Proteome Researchat the University of Liverpool for their invaluable advice on proteomictechniques and approaches and staff at the Philip Leverhulme EquineHospital, University of Liverpool, for their involvement with samplecollection.

Ethics

University of Liverpool Ethics approval ref.: VREC175.

Author contributions

Wrote the manuscript (J.A., M.P., A.S., L.R.M.), revised the manu-script (J.A., M.P., A.S., P.C., L.R.M.), analysed the data (J.A., D.S., A.S.,M.P.), experimental design (J.A., M.P., D.S., L.R.M.), collected clinicalsamples (L.R.M.). All authors read and approved the final manuscript.

Funding

The project was funded by a Veterinary School Research ProjectSupport Grant (VSRP VET032) and the Technology Directorate voucherscheme (www.liverpool.ac.uk/technology-directorate), University ofLiverpool. JA is funded by the Horse Trust. MJP is funded through aWellcome Trust Clinical Intermediate Fellowship. AS is funded by aKazakhstan President's Bolashak Scholarship.

Conflict of interest

The authors have no conflicts of interest.

Appendix A. Supplementary data

Supplementary data to this article can be found online at https://doi.org/10.1016/j.jprot.2019.04.020.

References

[1] L.M. Rubio-Martinez, C.R. Elmas, B. Black, G. Monteith, Clinical use of anti-microbial regional limb perfusion in horses: 174 cases (1999-2009), J. Am. Vet.Med. Assoc. 241 (2012) 1650–1658.

[2] K.T. Dicker, L.A. Gurski, S. Pradhan-Bhatt, R.L. Witt, M.C. Farach-Carson, X. Jia,Hyaluronan: a simple polysaccharide with diverse biological functions, ActaBiomater. 10 (2014) 1558–1570.

[3] C. Ruiz-Romero, F.J. Blanco, Proteomics role in the search for improved diagnosis,prognosis and treatment of osteoarthritis, Osteoarthr. Cartil. 18 (2010) 500–509.

[4] J. Mateos, L. Lourido, P. Fernandez-Puente, V. Calamia, C. Fernandez-Lopez,N. Oreiro, et al., Differential protein profiling of synovial fluid from rheumatoidarthritis and osteoarthritis patients using LC-MALDI TOF/TOF, J. Proteome 75(2012) 2869–2878.

[5] D.D. Frisbie, Synovial joint biology and pathobiology, in: Stick JAAaJA (Ed.),Equine Surgery, 3rd ed., Elsevier Inc, St. Louis, 2006, pp. 1036–1055.

[6] A.F. Sanchez-Teran, J.L. Bracamonte, S. Hendrick, L. Riddell, K. Musil, B. Hoff,et al., Effect of repeated through-and-through joint lavage on serum amyloid a insynovial fluid from healthy horses, Vet. J. 210 (2016) 30–33.

[7] S. Jacobsen, M.H. Thomsen, S. Nanni, Concentrations of serum amyloid a in serumand synovial fluid from healthy horses and horses with joint disease, Am. J. Vet.Res. 67 (2006) 1738–1742.

[8] M. Sinovich, N.F. Villarino, E.R. Singer, C.S. Robinson, L.M. Rubio-Martinez,Investiagtion of blood serum amyloid a concentrations in horses to differentiatesynovial sepsis from inflammation and determine prognosis an response to treat-ment, Vet. Surg. 47 (2018) E26.

[9] M.J. Peffers, B. McDermott, P.D. Clegg, C.M. Riggs, Comprehensive protein pro-filing of synovial fluid in osteoarthritis following protein equalization, Osteoarthr.Cartil. 23 (2015) 1204–1213.

[10] S. Jacobsen, T.A. Niewold, M. Halling-Thomsen, S. Nanni, E. Olsen, C. Lindegaard,et al., Serum amyloid a isoforms in serum and synovial fluid in horses with lipo-polysaccharide-induced arthritis, Vet. Immunol. Immunopathol. 110 (2006)325–330.

[11] C.S. Robinson, E.R. Singer, M. Piviani, L.M. Rubio-Martinez, Are serum amyloid Aor D-lactate useful to diagnose synovial contamination or sepsis in horses? Vet. Rec.181 (2017) 425.

[12] C.W. McIlwraith, D.D. Frisbie, C.E. Kawcak, C.J. Fuller, M. Hurtig, A. Cruz, TheOARSI histopathology initiative - recommendations for histological assessments ofosteoarthritis in the horse, Osteoarthr. Cartil. 18 (Suppl. 3) (2010) S93–105.

[13] J.A. Vizcaino, A. Csordas, N. del-Toro, J.A. Dianes, J. Griss, I. Lavidas, et al., 2016update of the PRIDE database and its related tools, Nucleic Acids Res. 44 (2016)D447–D456.

[14] A.H. Albar, H.A. Almehdar, V.N. Uversky, E.M. Redwan, Structural heterogeneityand multifunctionality of lactoferrin, Curr. Protein Pept. Sci. 15 (2014) 778–797.

[15] M. Paramasivam, K. Saravanan, K. Uma, S. Sharma, T.P. Singh, A. Srinivasan,Expression, purification, and characterization of equine lactoferrin in Pichia pas-toris, Protein Expr. Purif. 26 (2002) 28–34.

[16] Q. Ye, Y. Zheng, S. Fan, Z. Qin, N. Li, A. Tang, et al., Lactoferrin deficiency promotescolitis-associated colorectal dysplasia in mice, PLoS One 9 (2014) e103298.

[17] L. Edde, R.B. Hipolito, F.F. Hwang, D.R. Headon, R.A. Shalwitz, M.P. Sherman,Lactoferrin protects neonatal rats from gut-related systemic infection, Am. J.Physiol. Gastrointest. Liver Physiol. 281 (2001) G1140–G1150.

[18] P.H. Nibbering, E. Ravensbergen, M.M. Welling, L.A. van Berkel, P.H. van Berkel,E.K. Pauwels, et al., Human lactoferrin and peptides derived from its N terminus arehighly effective against infections with antibiotic-resistant bacteria, Infect. Immun.69 (2001) 1469–1476.

[19] M. Pammi, S.A. Abrams, Oral lactoferrin for the prevention of sepsis and necrotizingenterocolitis in preterm infants, Cochrane Database Syst. Rev. (2) (2015) 1–51(CD007137).

[20] P. Valenti, G. Antonini, Lactoferrin: an important host defence against microbialand viral attack, Cell. Mol. Life Sci. 62 (2005) 2576–2587.

[21] A. Afeltra, D. Caccavo, G.M. Ferri, M.A. Addessi, F.G. De Rosa, A. Amoroso, et al.,


9

http://www.liverpool.ac.uk/technology-directoratehttps://doi.org/10.1016/j.jprot.2019.04.020https://doi.org/10.1016/j.jprot.2019.04.020http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0005http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0005http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0005http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0010http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0010http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0010http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0015http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0015http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0020http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0020http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0020http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0020http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0025http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0025http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0030http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0030http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0030http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0035http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0035http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0035http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0040http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0040http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0040http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0040http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0045http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0045http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0045http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0050http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0050http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0050http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0050http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0055http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0055http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0055http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0060http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0060http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0060http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0065http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0065http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0065http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0070http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0070http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0075http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0075http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0075http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0080http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0080http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0085http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0085http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0085http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0090http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0090http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0090http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0090http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0095http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0095http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0095http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0100http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0100http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0105

Expression of lactoferrin on human granulocytes: analysis with polyclonal andmonoclonal antibodies, Clin. Exp. Immunol. 109 (1997) 279–285.

[22] R.M. Bennett, J.L. Skosey, Lactoferrin and lysozyme levels in synovial fluid: dif-ferential indices of articular inflammation and degradation, Arthritis Rheum. 20(1977) 84–90.

[23] J.J. Abbink, A.M. Kamp, E.J. Nieuwenhuys, J.H. Nuijens, A.J. Swaak, C.E. Hack,Predominant role of neutrophils in the inactivation of alpha 2-macroglobulin inarthritic joints, Arthritis Rheum. 34 (1991) 1139–1150.

[24] D. Caccavo, P. Garzia, G.D. Sebastiani, G.M. Ferri, S. Galluzzo, M. Vadacca, et al.,Expression of lactoferrin on neutrophil granulocytes from synovial fluid and per-ipheral blood of patients with rheumatoid arthritis, J. Rheumatol. 30 (2003)220–224.

[25] Decoteau E, Yurchak AM, Partridge RE, Tomasi TB, Jr. Lactoferrin in synovial fluidof patients with inflammatory arthritis. Arthritis Rheum. 1972;15:324–5.

[26] C. Zhang, Y. Li, W. Tang, N. Kamiya, H. Kim, Lactoferrin activates BMP7 geneexpression through the mitogen-activated protein kinase ERK pathway in articularcartilage, Biochem. Biophys. Res. Commun. 431 (2013) 31–35.

[27] M.E. Shirtliff, J.T. Mader, Acute septic arthritis, Clin. Microbiol. Rev. 15 (2002)527–544.

[28] S. Baveye, E. Elass, J. Mazurier, G. Spik, D. Legrand, Lactoferrin: a multifunctionalglycoprotein involved in the modulation of the inflammatory process, Clin. Chem.Lab. Med. 37 (1999) 281–286.

[29] R.D. Baynes, W.R. Bezwoda, Lactoferrin and the inflammatory response, Adv. Exp.Med. Biol. 357 (1994) 133–141.

[30] M.L. Kruzel, Y. Harari, D. Mailman, J.K. Actor, M. Zimecki, Differential effects ofprophylactic, concurrent and therapeutic lactoferrin treatment on LPS-induced in-flammatory responses in mice, Clin. Exp. Immunol. 130 (2002) 25–31.

[31] A. Alunno, O. Bistoni, M. Manetti, G. Cafaro, V. Valentini, E. Bartoloni, et al.,Insulin-like growth factor binding protein 6 in rheumatoid arthritis: a possible novelchemotactic factor? Front. Immunol. 8 (2017) 554.

[32] D. Aboalola, V.K.M. Han, Different effects of insulin-like growth Factor-1 and in-sulin-like growth Factor-2 on myogenic differentiation of human mesenchymalstem cells, Stem Cells Int. 2017 (2017) 8286248.

[33] O. Nedic, D. Robajac, M. Sunderic, G. Miljus, B. Dukanovic, V. Malenkovic,Detection and identification of oxidized insulin-like growth factor-binding proteinsand receptors in patients with colorectal carcinoma, Free Radic. Biol. Med. 65(2013) 1195–1200.

[34] S. Lu, S. Purohit, A. Sharma, W. Zhi, M. He, Y. Wang, et al., Serum insulin-likegrowth factor binding protein 6 (IGFBP6) is increased in patients with type 1

diabetes and its complications, Int. J. Clin. Exp. Med. 5 (2012) 229–237.[35] J.R. Anderson, M.M. Phelan, P.D. Clegg, M.J. Peffers, L.M. Rubio-Martinez,

Synovial fluid metabolites differentiate between septic and nonseptic joint pathol-ogies, J. Proteome Res. 17 (2018) 2735–2743.

[36] M. Levi, T.T. Keller, E. van Gorp, H. ten Cate, Infection and inflammation and thecoagulation system, Cardiovasc. Res. 60 (2003) 26–39.

[37] A.K. So, P.A. Varisco, B. Kemkes-Matthes, C. Herkenne-Morard, V. Chobaz-Peclat,J.C. Gerster, et al., Arthritis is linked to local and systemic activation of coagulationand fibrinolysis pathways, J. Thromb. Haemost. 1 (2003) 2510–2515.

[38] S.M. Andreassen, A.M.L. Vinther, S.S. Nielsen, P.H. Andersen, A. Tnibar,A.T. Kristensen, et al., Changes in concentrations of haemostatic and inflammatorybiomarkers in synovial fluid after intra-articular injection of lipopolysaccharide inhorses, BMC Vet. Res. 13 (2017) 182.

[39] S. Jacobsen, M. Kjelgaard-Hansen, Evaluation of a commercially available appa-ratus for measuring the acute phase protein serum amyloid a in horses, Vet. Rec.163 (2008) 327–330.

[40] A.M. Blom, The role of complement inhibitors beyond controlling inflammation, J.Intern. Med. 282 (2017) 116–128.

[41] E. Sakiniene, T. Bremell, A. Tarkowski, Complement depletion aggravatesStaphylococcus aureus septicaemia and septic arthritis, Clin. Exp. Immunol. 115(1999) 95–102.

[42] S.E. Olsson, S. Reiland, The nature of osteochondrosis in animals. Summary andconclusions with comparative aspects on osteochondritis dissecans in man, ActaRadiol. Suppl. 358 (1978) 299–306.

[43] Y. Zhang, F. Vasheghani, Y.H. Li, M. Blati, K. Simeone, H. Fahmi, et al., Cartilage-specific deletion of mTOR upregulates autophagy and protects mice from osteoar-thritis, Ann. Rheum. Dis. 74 (2015) 1432–1440.

[44] D. Cejka, S. Hayer, B. Niederreiter, W. Sieghart, T. Fuereder, J. Zwerina, et al.,Mammalian target of rapamycin signaling is crucial for joint destruction in ex-perimental arthritis and is activated in osteoclasts from patients with rheumatoidarthritis, Arthritis Rheum. 62 (2010) 2294–2302.

[45] N. Maicas, M.L. Ferrandiz, R. Brines, L. Ibanez, A. Cuadrado, M.I. Koenders, et al.,Deficiency of Nrf2 accelerates the effector phase of arthritis and aggravates jointdisease, Antioxid. Redox Signal. 15 (2011) 889–901.

[46] E. Chiaradia, M. Pepe, M. Tartaglia, F. Scoppetta, C. D'Ambrosio, G. Renzone,L. Avellini, F. Moriconi, A. Gaiti, A. Bertuglia, F. Beccati, A. Scaloni, Gambling onputative biomarkers of osteoarthritis and osteochondrosis by equine synovial fluidproteomics, J. Proteomics 75 (14) (2012) 4478–4493.


10

http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0105http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0105http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0110http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0110http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0110http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0115http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0115http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0115http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0120http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0120http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0120http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0120http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0125http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0125http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0125http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0130http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0130http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0135http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0135http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0135http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0140http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0140http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0145http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0145http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0145http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0150http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0150http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0150http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0155http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0155http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0155http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0160http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0160http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0160http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0160http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0165http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0165http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0165http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0170http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0170http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0170http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0175http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0175http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0180http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0180http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0180http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0185http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0185http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0185http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0185http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0190http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0190http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0190http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0195http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0195http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0200http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0200http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0200http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0205http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0205http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0205http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0210http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0210http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0210http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0215http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0215http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0215http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0215http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0220http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0220http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0220http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0225http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0225http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0225http://refhub.elsevier.com/S1874-3919(19)30133-2/rf0225

The synovial fluid proteome differentiates between septic and nonseptic articular pathologiesIntroductionMethodsSample collectionSynovial fluid preparationLiquid chromatography tandem mass spectrometry (LC-MS/MS) and label-free quantificationPathway and network analysis of proteomic data in contrasts of either sepsis, OA or OC versus healthyValidation of mass spectrometry using ELISAAdditional statistical analyses

ResultsProtein concentration of SF different arthropathiesDifferential abundance of proteins in healthy compared to osteoarthritis, osteochondrosis and sepsis SFDE proteins were filtered to include proteins identified with two or more peptides, q  2 fold change in expression

DiscussionConclusionAcknowledgementsmk:H1_17Ethicsmk:H1_19Author contributionsmk:H1_21Fundingmk:H1_23mk:H1_24Supplementary dataReferences

Journal of Proteomics - University of Liverpool · Proteomics encompasses the comprehensive...

Documents

Transcript of Journal of Proteomics - University of Liverpool · Proteomics encompasses the comprehensive...